Imagine a U of Florida Halloween party in the early 1990s, at the affectionately-named and unfortunately-rank Zoohaus, with all of us Zoology and Botany grad students costumed in our favorite evolutionary phenomena. Wild Type, Kamikaze Sperm, Plastic Phenotype … you can guess plenty of others. I put on a pair of tan short-shorts, cut the lower legs off a pair of old jeans and tied them onto my legs, and found an old trench coat. With socks and sneakers, it looked like I was wearing pants under the coat, and I added a disreputable-looking hat. So, a flasher costume … and on the front of the shorts, I taped a sign saying “Adaptation,” and on the back of the trench coat, a sign saying “Constraint.”
These were fightin’ words in evolutionary theory twenty to twenty-five years ago, and to a lesser extent they can be today. Not because anyone disagrees with what the above cartoon shows – raw physics rule, after all, and no one expects natural selection to subvert it. What gets biologists all hinky is when other biologists suggest the presence of historically-contingent genetic/physiological constraints, which don’t have to be there but are. For instance, the four legs of tetrapods. You can lose’em, but no one in this group has ever added any. Or to pick more-or-less at random, the absence of horns in carnivoran mammals – the textbook view must say, “gee I guess they never ‘needed’ them” or more likely it changes the subject.
What upsets the conversation is that selection is – romantically speaking – supposed to be able to do anything the given organism “needs,” or which “matches” them to a given environment. As far as I can tell, the only justification for this perspective is that we, as human observers, find the result beautiful and we want to call the current array of diversity a system of some kind, shaped by Nature. Talking about what such a system doesn’t do is … I guess, rude.
In my generation of biologists, there’s a bit of a myth that “all was natural selection” until the 1970s and then “mavericks” came along and caused trouble with this talk of constraint and similar matters. However, I have often found the pre-1950s biologists to be much more inventive and critical than their locked-down successors during a couple critical decades – the ones which produced the standard teaching texts, unfortunately. The older taxonomists were surprisingly cladistic in their thinking, for instance, and they saw famous and vocal figures as George Gaylord Simpson and Ernst Mayr as mere colleagues rather than demigods and thought-leaders.
It was my good fortune to be mentored immediately after college, pre-grad school, by perhaps the last, certainly one of the greatest examples: Philip Hershkovitz, for whom a brief biography is here and a more extensive one here. I can’t improve upon Bruce Patterson’s review in the latter link [it’s from the 1987 special issue or Festschrift honoring Hershkovitz in Fieldiana (Zoology), volume 39], but in a word, he was one of the most significant contributors to our understanding of mammalian diversity in the topic’s history. Hershkovitz was also almost unique in the second half of the twentieth century in studying primates as mammals as opposed to proxy humans or “cousins.” (the only other example I can think of was the kind and thoughtful Jack Fooden, whose office was down the hall)
He was also uniquely immune to professional status pressure, partly due to his unusual Research Curator position without university affiliation and its concerns about promotion, partly due to not needing to truckle to journal editors, and partly because he was – and no offense intended – one of the toughest old bastards I have met in a lifetime of encountering such. All three of these features had been turned up to 11 since his official if otherwise nonexistent retirement in the 1970s.
His concept of metachromism – my topic here – dates from the late 1960s, but it has not penetrated far into evolutionary theory nor into standard ecological or systematic texts. It may be asking too much to accept me as a worthy posthumous voice in this issue, but I did assist him throughout the preparation of his classic revision of Callicebus, including manuscript review and intensive measurement-and-biogeography sessions, and I do claim personal contact with his most considered views and applications of the idea. Callow I may have been, but he didn’t fire me as he had so many others, and along with his
extensive descriptions observations of my shortcomings, he was right in saying I was soaking it all in, and that I was thinking. For what it’s worth, I haven’t stopped doing that.
I’ll begin with a brief look at the well-known fundamentals of mammalian hair pigmentation. There are two pigments: eumelanin (brown) and phaeomelanin (red). The most common distribution is observed in alternating bands, generating a sandy or “rich” look to fur color, called agouti.
Variations in mammal hair color result from a change in the banding. As depicted below, the simplest version is altered widths for the bands in any possible variant, sometimes to the extent of the hair becoming unbanded and therefore featuring pigment throughout its length. Human hair, for example, is all unbanded.
However, this only begins the nuances of possible modification, because the hair might remain banded but lose one pigment, generating alternating bands of white and pigmented; because the two pigments can occur in the same band or whole hair, for an auburn result; and because the density of pigmentation is its own independent variable, so that the same banding pattern can look very different on that basis alone (up to and including fully unpigmented, white hair).
As I implied, none of the above is any sort of controversy. Hershkovitz proposed, however, that (i) the designated changes may occur without selective pressure, and (ii) alteration from agouti and loss of pigmentation are irreversible. Once such a change occurred, subsequent alterations proceeded from that one, whether further “unbanding” or pigment loss, and didn’t go back to agouti. There are three distinct stages of logic in such a claim.
- Perceiving the changes from agouti as a constraint, from observing only certain changes as a pattern throughout phylogenies
- Proposing that these changes arise from a process with an unspecified but intrinsic reliability, independent of other known processes such as selection
- Applying it, sometimes, as a device in constructing phylogenies, especially to resolve very closely related species and subgroups for which other variables were too coarse-grained
During the 1960s, in which the intellectual context for evolution was literally defined and even socio-professionally ruled by Simpson and Mayr, such talk was raw heresy. The word was that mutations (variation) may be random, but selection was conceived as a highly-determinant boon to a given organism, associated with such terms as optimality, maximum advantage, and especially niche. This outlook was rooted less in textual Charles Darwin than in Alfred Russell Wallace’s views, themselves based on the concepts of ecosystems as harmonious complexes and of selection which could necessarily “go” wherever it “needed.” Furthermore, systematics itself was entering a state of flux in which this very topic – directionality of change – was a particularly sore point.
I haven’t been impressed by the historical critique, with the exception of Timothy Lawlor’s initial response in Evolution in 1969. Most of it since then, e.g. Lee Howarth’s Genesis 2.0, looks more like pushback than rebuttal. If one wanted to rebut, the obvious first step would be to call for validation of the directionality at the pattern level, using, for lack of a better word, third-party consultation. In other words, find a critter (genus) with (i) lots of species with varying and characteristic modifications of the agouti pattern and (ii) a well-resolved and generally noncontroversial phylogeny which uses variables (“characters”) that aren’t fur-color based, and then see if the hair-pattern changes refute Hershkovitz’s proposed constraint. As Patterson has pointed out more than once, this could have been done easily at any time since 1968 but to date, no takers. Then, pending the pattern showing up as valid preferably in more than one group, one might make use of all this fancy gene-expression tech we have available now to examine the process – for instance, whether agouti coloration can be “knocked in” through genetic manipulation, which would be remarkably easy to validate with RNA arrays and to test in lab mice, as such techniques go. Studies like these would set up the appropriate parameters, if any, for applying the principle to systematics. Or even better, to examining the concept of inherited constraint – without which, as I see it, all talk of selection of “doing” this or “doing” that is dancing on air. Pattern – process – application; this is science 101 and I shouldn’t have to be pointing that out.
Instead, unfortunately, metachromism has the reputation of “Hershkovitz’s weird old thing,” and the few authors who’ve mentioned it focus instead on the red herring of disavowing it as a magic wand for determining phylogeny. However, Hershkovitz never used it that way nor claimed it should be so used; neither did he claim that selection had nothing to do with coat color or that metachromism would consistently gallop past selection. These straw men are pretty irritating to read actually, as if “but then all mammals would be white by now” or “but color does often match environment” were supposed to be rebuttals.
Granted, Hershkovitz took the idea along this pattern-to-application path awfully fast, black-boxing process along the way, and also all by himself, which is arguably high-handed. Further, in responding to feedback, he stopped barely short of flipping the bird to anyone whom he considered less than rigorous … “extemporaneous and unconventional,” “acerbic,” and “something other than conciliatory or diplomatic” are Patterson’s phrases regarding his characteristic approach to dialogue, for which the targets of his wrath probably had their own terms. Still, that no one picked up the question as a serious research item on its own merits is surprising.
To the point, then: that even at this late date almost fifty years since he proposed the idea, and almost twenty since his death, the intellectual circumstances which permit/don’t permit addressing it have barely changed.
Next: Fitter than thou